Data sharing in neuroscience

Data sharing is one of most important challenge in several scientific fields and it has already given good results in genomic and proteomic analysis. Also neuroscience should benefit from data sharing, in particular from functional magnetic resonance imaging (fMRI) data sharing. This technique is used to map functional region of brain that are monitored and recorded in resting or working state. The entire map of functionally tuned regions in the brain constitutes the connectome and it may help scientists to better understand brain physiology and functioning. Michael Milham and his group of psychiatry at the New York University propose to other scientists worldwide to share their images of functional magnetic resonance in order to dispose more samples to analyze and compare. Indeed, even if these images have been generated in uncoordinated manner and with different purposes, data sharing could be extremely useful. Thus, they proposed the 1000 Functional Connectomes Project in which they wanted to collect data from fMRI. More than 35 centers decided to have part in this project, shared the images of more than 1400 volunteers, and deposited them into an open access database. Applying analytic and computational methods to evaluate and aggregate shared data, Milham and coworkers demonstrated a universal architecture of functional connections in the resting human brain. Furthermore, they highlighted the consistent loci of variability between individuals and centers and the brain region for which age and gender emerged as significant determinants. This study was published on PNAS Journal in 2010. Other follow-up studies will be necessary to validate the results described here. For instance, the information collected in the Functional Connectomes database could be used to build normative maps of functional system in the brain, as well as to interpret laboratory test results, thus using this method for a clinical application. The Functional Connectomes project is continuously growing and their supporters and creators encourage all groups involved in neuroscience imaging to share their data. This collaborative effort will give important scientific result because of the synergistic effect of data sharing: indeed, data that are brought together can be much more than the sum of the parts. Based on that, data sharing could be crucial also in other medical disciplines where imaging may explain biological processes. Of course, specific software will be necessary to effectively manage the huge amount of data derived from data sharing and the computational biologist will be fundamental in the laboratory. In conclusion, sharing data and computational analysis have become milestones in several field of science, and we are sure that this collaboration will give a significant contribution in scientific advances.

Novel nomenclature system in botanic

The general rule to assign a name to novel plant species is to publish the name self on printed journal, in order to guarantee the immutability of the publication. In the modern era, electronic journals make possible a faster data sharing between scientists and also botanic must use these systems. Thus, for the first time four novel plants have been named and described in an article appeared in the Plos One journal, available only on-line.
The, authors printed 10 copies and distributed them to some libraries around the world in order to correctly follow the rule of the international code of botanical nomenclature and get their paper published. Even if the overall story is not so revolutionary, for the first time it has been possible a tempestive communication of data to other scientists. Furthermore, all scientists have been able to reach this paper, because it’s more probable that institutes buy an electronic journal rather the printed one. After the botanical example, also the international commission of zoological nomenclature started an updating process and positively considered the opportunity to open to the electronic journals. In all cases, we are sure that “verba volant, scripta manent” and printed copies allowed us reading experiments performed hundreds years ago.

The value of negative results

Negative results are similarly important than positive results for science advances. Indeed, a large number if hypothesis could be excluded if we considered negative results from other studies. Unfortunately, negative results are not so easily accepted and published in peer reviewed journal, thus are not available for science community.
Scientists from the German Research Centre for Environmental Health and from the University of Munich spent their time to collect negative results from protein-protein interaction studies and built a database, called “negatome”. They used data from literature search and from structural information retrieved into Protein Data Bank and identified almost two thousand of not interacting pairs. The novelty is that the not-interaction is experimentally demonstrated and published, while in previous work not- interaction was determined from not co-localization: if two proteins are differentially located, they will not interact. This sentence could be true, but is not predictive of interaction, just co-localization. The Negatome can contain some false negative pairs, but in general it can be considered a useful tool to compare and confirm results from immunoprecipitation or two hybrid system or other techniques that usually generate some false positive. We hope that this example should be followed by other databases for negative results.

Statistics in science

A common English quote says that there are three kinds of lies: lies, damned lies and statistics. This sentence was wrongly attributed to Disraeli, whereas it was firstly published by Marc Twain. Even if it could seem true, statistics is the only tool available up to date to manage a large amount of data and to give their a mean. Several functions are usually used in statistical analysis; the most common are the average, standard deviation, frequency and also the number of observation is reported in scientific publications.
All these data are defined as descriptive statistics because of their direct characterization of the population of interest. In other cases more complex calculations must be done to find significance from the study. Linear or logistic regressions are models to identify the trend that better explain and summarize the collected data. From linear or logistic regression is possible to interpolate or extrapolate data in order to predict results into the studied range or outside the studied range, respectively. Furthermore, from statistical analyses prevalence, incidence, absolute risk, odd ratio and relative risk are determined. All of these parameters are so important that must be determined during clinical study and are subsequently used to compare different protocols, drugs and so on. Clinical studies are classified as experimental studies or observational studies. The main difference between these two classes is the presence of a treatment given to patients in the experimental studies, whereas in the observational studies none treatment is followed but the selected population is just observed to determine the parameters of interest. In both cases, given the high number of patients enrolled in order to obtain significant results, statistics is useful to compare treated cohort and placebo cohort, for example, as well as to evaluate the role of specific component of the study, namely called covariates.
The major part of software that help to collect and manage data from a clinical study have also some algorithms to calculate standard statistical parameters during the data collection self. Revision and further elaboration must be done by professionals in order to correctly consider study results. Statistics is important not only in clinical trials, but also in all experiments performed in science. Indeed, to be sure to have obtained a result as a consequence of certain conditions and not due to serendipity, all experiments are usually repeated three or more times. Data presentation normally comprises average and standard deviation or confidence interval and significance is determined by a series of tests that should be described in material and methods section of the publication. In conclusion scientists must have some basis of statistics because this discipline confers value to their experiments and make them shareable and comparable with scientific community. Even if statistics is considered as a lie in some cases, it will be useful for science advances.

IT system protection

The broad diffusion of technologies such as computers, databases, mobile phones and so on, has determined a considerable change in our life. We usually use this kind of instruments at the work or school, at home and in other situations of everyday life.
It’s really important that the global IT system would be safe and protected from hacking attacks. Last year, an important mobile phone manufacturer company found vulnerability in its system and phone calls could be listen by everyone knew this failure. This first episode started an intensive research on security system in order to preserve IT and our privacy. Scientists were called to identify flaws into algorithms, acting like legal hackers, and were invited to share their skills and experiences with the community. So, manufactures changed their habits to reject the possibility to reveal flaws in their systems and choose an ethic approach to improve the worldwide knowledge. Also manufacturers that produce IT systems for pharmaceutical companies must consider the opportunity to look for new devices to protect all sensible data regarding the clinical study. Indeed, personal details of patients contained in clinical study are usually shared between centres involved in the trial and become more sensitive of possible attacks. Thus, it’s ethic that these data should be preserved.

Biological novel… Why not?

In the last number of Nature Methods, the editorial talks about a curious topic: biology in novel. Biologists become characters and their dilemmas, scientific and ethical problems and every-day stories are the main topics of numerous books; perhaps, this strange and interesting kind of literature could help to destroy the idea of biologist like a crazy scientist or a mouse always closed in his lab-cage.

Everyone that works in lab has a lot of amazing stories to tell and intrigues and mysteries are often associated to publications, specially in really famous journals. Love and jealousy arise between benches, as well as data manipulation to obtain Nobel prize results, crucial ethical problem to solve before starting some experiment or intense intellectual relationship are often mentioned during scientist discussion: all these aspects of scientific life could become obvious topics for a “biological” novel. Some times, reading about ourselves could have a cathartic effect: so, during the summer it could be interesting for scientists reading these books in order to acquire an external point of view of their lives and, maybe, enjoy for all the problems that usually encounter in their work. So good reading! The website where you can find these books is www.lablit.com.

Quality control in lab

Quality control in laboratory is essential to obtain great results. Several studies demonstrated that more publications are accepted by peer reviewed journals when good laboratory procedures are followed. This means that giving the same time and the same amount of money, labs that work with GLP are more productive than those that aren’t in GLP. Good laboratory procedures are rules (SOPs) which scientists have to keep in mind when they perform experiments and have to be the same for all people working in the lab.

Standard operating procedures (SOPs) should be written for each instrumentation present in lab and represent guidelines to work with. From calibration to final cleaning, terms of use of an instrument are well described in order to guarantee firstly that instrument is correctly used, secondarily that everyone in lab uses machine in the same manner: this is a crucial point to allow comparison of results produced in lab. A training has to be done before using an instrument, at least the first time with a manufacturer’s specialist and then by the most expert person in lab. Each instrument has a responsible that takes care of management and maintenance. Ideally, in this way measurements obtained from an instrument by all users are comparable and consistent and standard deviation between repeated experiments should decrease. SOPs are not applied only to instruments use, but can also describe other important actions, normally performed in lab. For instance, when data management has well defined rules, it’s easier and faster to retrieve information. Whatever kind of data elaborated in files –texts, tables, pictures, graphs- should be classified and named with a code that, for instance, contains project number, operator, day, kind of file: so just reading the name, it should be possible to understand if we have found what we are looking for. Standardization of data management is crucial when there are a lot of people in lab or there is a quick turnover, indeed these are common cases in which some data could be lost. Writing a notebook is another part of scientific work that is essential for scientist and also this aspect has to be standardized. Indeed, in notebook protocols are usually described, raw data collected, first observations noted and each scientist tends to personalize his book.
A standardized manner of writing notebook is important to make easier sharing protocols and ideas between group members. Last but not least, all reagents from salts to enzymes, from culture media to animals, have to be registered in terms of availability, arrival date, expiration date in a common repository accessible to all people working. In this way, scientists can quickly check to have all reagents for their experiment before starting and don’t waste time and, more important, they don’t waste money to buy reagents maybe already available in lab, but hidden in some dark corner. Software is available to correctly manage all kind of scientific repository, from freezers to nitrogen tank to collect cells. In conclusion, following GLP makes experimental work more efficient and less expensive.

The Activity based protein profiling

About 30% of human proteins have not been characterized yet in terms of activity and substrates, but their incorrect functionality is often one molecular basis for human diseases.
A research group of The Scripps Research Institute tried to overcome this problem and developed a system, usable in high throughput screening, to analyse not well characterized proteins.

Their ABPP (Activity based protein profiling) approach uses a chemical probe enable to fit in catalytic site of a broader range of proteins, belonging to the same family. In this way it is negligible to know all biochemical properties of the specific protein or the ideal substrate for each target. The probe is conjugated to a polarized fluorophore that could monitor the change in the catalytic site conformation in presence of the inhibitor and make possible to use this technique in high throughput screening.
One limitation of this method is the need, at least in high throughput, for purified proteins, otherwise with such a generic substrate it could be difficult to relate results of a screening to a specific protein. In this case it’s crucial to retest positive hits in gel based assays in order to obtain straightforward data. However, ABPP represents also a good starting point for biochemical characterization of unknown proteins.

Small molecule inhibitors and resistance

A great challenge in academic labs and pharmaceutical industries is the identification of small molecule inhibitors that could specifically target proteins involved in human diseases.
To do this, the most important step is the knowledge of biochemical features of the target protein. Human diseases associated to the target are usually well understood and the target has been previously validated: this means that its inhibition determines a recovery of normal cellular functionality.

X-ray crystallographic analysis and SAR (structure activity relationship) studies are crucial to identify and improve small molecules; otherwise, high throughput screening are performed, this route is usually followed by pharmaceutical or biotech companies that dispose of instrumentation to run the screening and manage data. Several small molecules identified as potential drug in one disease are tested in clinical studies in order to bypass resistance problems that could arise during long term treatment. Indeed, during this therapies a natural selection of mutations that provide the restoring of protein target activity, happens and makes useless the treatment.
Nevertheless, it’s necessary to continue research of small molecule inhibitor, especially in view of good results obtained, for instance, in tumoral diseases.